Tom Maguire had a good comment on my last post about Folksonomies in the Enterprise. He said (to paraphrase) that what makes folksonomies more powerful than taxonomies is that the "authority" of knowledge and control is different. I totally agree, and wanted to give a real-world example.
Look at how words come into the language... fundamentally, people make them up. If they are useful, interesting, catchy, or just plain lucky, then other people start using the same word too. Sometimes the word gets used to mean slightly different things, but eventually the set of uses settles down. Then, along comes Merriam-Webster (and others like them) whose role in life is, get this, to write it down - not to invent it, not to create it, not to define it, just to document it.
Sometimes, a word changes in meaning over time. Again, this change isn't initiated by Merriam-Webster - they just document it. This process of language evolution has worked well for centuries and has probably helped speed up with the advent of the internet - the lag from word adoption to documentation has shortened.
Who's in charge of the language? Notice that it's not Merriam-Webster. Languages evolve just like folksonomies. Terms get invented, they get popular, then they settle into common use. Who's in charge? Folks are in charge. Power to the people!
Taxonomies, on the other hand, are the exact opposite of this. A central authority deems itself the only one capable of defining "proper" terminology. My favorite example is the French language authority. If, for example, you want to wish someone a good weekend the authority has deemed that the correct way is to say, "Bon fin de semaine" ("good end of week" translated literally). How do the majority of French people wish you a good weekend? "Bon fin de semaine"? Nope, "Bon weekend." Has the central authority caught up with (or even acknowledged) common use? I'll let you guess...
SOA What does all of this mean? There are a few important differences between the approaches:
- Folksonomies naturally evolve (they survive changes in meaning). But, on the other hand, it's like pulling teeth to evolve a taxonomy.
- Folksonomies don't have to be "right" - precisely because they can evolve easily. This means you can start small and expand, versus a taxonomy approach where you need to be perfect upfront or face the pain of changing it later.
So, back to the topic of my original post. If you have a small SOA architecture team, a central authority (taxonomy based approach) can work, but this doesn't scale to a large team. On the other end, a folksonomy based approach works well with small or large teams. If only SOA governance repositories actually supported folksonomies!